Gary Mills wrote:
On Fri, Nov 02, 2007 at 04:16:14PM -0400, Bob Doolittle wrote:
But I don't know how NAT would come into play. There's little
in common with logging into Solaris and logging into Windows
with uttsc. Perhaps larger WAN latencies and/or a lossy WAN
link exacerbate the problem of heartbeat starvation under paging
load. What does utcapture tell you during this period? How about
vmstat, do you get a spike in paging activity during this period?
I disagree. Both logging in and starting a Windows connector session
initiate a new connection to the Sun Ray client. NAT means that my
Sun Ray can initiate connections, but can't accept them without
cooperation with the NAT router. I included the snoop results in a
previous message.
I didn't see any snoop output. Only an
interpretation that at some point UDP traffic
stopped and you got some ICMP Destination
Unreachables from the *server*. The details of
that ICMP message may be illuminating, as well as
what immediately preceded the problem. Is your X
server dying? It's possible that some process on
the server is dying and hence no app is reading
that port any more.
It's pretty clear. Something changed in the most
recent version of the Sun Ray server software to provoke this
behavior.
Here are some architectural details that might help you
understand SRSS better:
- uttsc doesn't create new connections to the
client. It simply is an X client that connects
to windows using RDP, and uses the existing
client UDP "connections" for X rendering
traffic.
- The only exception to this is that the wrapper
starts up a new, dedicated instance of utaudio
to handle the audio stream. utaudio, like the X
server and now device services, uses a special
"CALLME" protocol to handle NAT, where the
existing 7009/tcp connection is used to tell the
client to send a UDP packet in order to initiate
a new "connection", in order to handle punching
through a NAT firewall. Once a UDP packet has
been sent outwards to a particular server port
from a particular client port, the NAT firewall
ought to allow incoming traffic between the same
ports.
I'm using VPN from behind a NAT firewall and
can connect to Windows with uttsc without problems
with SRSS 4.0, so whatever you're experiencing is
not ubiquitous.
It's interesting that it's the server sending the ICMP
packets. That doesn't sound like an obvious NAT
problem with the client.
-Bob
Disclaimer: Opinions expressed in this mail are my own,
and are not necessarily shared by my employer
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users
